- Email: [email protected]
Differences in Cytotoxicity of Native and Engineered RIPs Can Be Used to Assess Their Ability to Reach the Cytoplasm
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS ARTICLE NO.
249, 637–642 (1998)
RC989207
Differences in Cytotoxicity of Native and Engineered RIPs Can Be Used to Assess Their Ability to Reach the Cytoplasm Maria Svinth,* Jo¨rg Steighardt,* Raquel Hernandez,* Jung-Keun Suh,* Curtis Kelly,* Philip Day,† Michael Lord,† Tomas Girbes,‡ and Jon D. Robertus*,1 *Institute of Cellular and Molecular Biology, Department of Chemistry and Biochemistry, University of Texas, Austin, Texas 78712; †Department of Biological Sciences, University of Warwick, Coventry CV4 7AL, England; and ‡Departmento de Bioquimica y Biologia Molecular, Facultad de Ciencias, Universidad de Valladolid, 47005 Valladolid, Spain
Received July 20, 1998
Ricin is a heterodimeric cytotoxin composed of RTB, a galactose binding lectin, and RTA, an enzymatic Nglycosidase. The toxin is endocytosed, and after intracellular routing, RTA is translocated to the cytoplasm where it inactivates ribosomes resulting in a loss of host cell protein synthesis and cell death. We show for the first time that the cytotoxicity against cultured T cells by several RTA mutants is directly proportional to the enzyme activity of RTA, suggesting this is a reliable system to measure translocation effects. Large discrepancies between cytotoxicity and enzyme action for a given pair of toxins are therefore attributable to differences in cell binding, uptake, or membrane translocation. Fluid phase uptake and cytotoxicity of isolated RTA are essentially identical to that of the single chain toxin PAP. This important finding suggests that RTA, and the A chain of class 2 RIPs in general, has not evolved special translocation signals to complement the increased target cell binding facilitated by RTB. Experiments with the lectin RCA and with ebulin suggest those toxins have diminished cytotoxicity probably mediated by comparative deficiencies in B chain binding. Addition of a KDEL sequence to RTA increases fluid phase uptake, consistent with the notion that transport to the ER is important for cytotoxicity. Fusion of MBP or GST to the amino terminus of RTA has little effect on enzyme action or cytotoxicity. This result is not altered by protease inhibitors, suggesting the fusion proteins are probably not cleaved prior to translocation of the toxic A chain and implying that the toxins can carry large passenger proteins into the cytoplasm, an observation with interesting potential for analytical and therapeutic chemistry. q 1998 Academic Press
1 To whom correspondence should be addressed. Fax: (512) 4716135. E-mail: [email protected].
Ricin is a 65 kD heterodimeric cytotoxin composed of an A chain (RTA) and B chain (RTB) linked by a disulfide bridge (1). RTA is an N-glycosidase which recognizes a specific adenine within 28S rRNA (2), inactivating protein synthesis. RTA is a member of the ribosome inactivating proteins (RIPs) which also includes a number of single chain RIPs, such as pokeweed antiviral protein (PAP), trichosanthin, and gelonin (3). These proteins are all of similar size and all carry out the same N-glycosidation reaction (4). RTB is a lectin with an affinity for galactose; the bound toxin is endocytosed and may then traffic within a variety of intracellular compartments (5). RTA is released from RTB by reduction of the disulfide bond, and eventually reaches the cytoplasm where it enzymatically attacks ribosomes and kills the target cell (6). A good deal is known about the biochemistry and structure of RTA (7,8). Site directed mutagenesis (9,10,11) and X-ray analysis of substrate analogs have led to a mechanism of action for the RTA depurination reaction (12). RTB is a two-domain protein with two galactose binding sites (13). Site-directed mutations have confirmed the importance of sugar binding residues deduced from the structure (14). Details concerning toxin uptake, processing, intracellular routing, and membrane translocation are much less clear (15). Ricin is so toxic that only one, or a very few, molecules within the cytoplasm are required to kill a target cell. As a consequence, it is difficult to distinguish the fate of the intoxicating molecules from the bulk trafficking of what may be nonproductive pathways. On the other hand, reduction in cellular protein synthesis is an extremely sensitive indicator that RTA has reached the cytoplasm. As a consequence, protein synthesis may be used as an indirect measure of translocation of toxins into the cytoplasm.
637
0006-291X/98 $25.00 Copyright q 1998 by Academic Press All rights of reproduction in any form reserved.
AID
BBRC 9207
/
695c$$$181
08-12-98 21:55:10
bbrcg
AP: BBRC
Vol. 249, No. 3, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
Lord and coworkers have proposed that after cellular uptake, RTA may be routed to the endoplasmic reticulum (ER) and escape from there into the cytoplasm (16). This hypothesis is supported by their observation that attaching the tetrapeptide KDEL (Lys-Asp-GluLeu) to the C terminus of RTA (RTArKDEL) increases cytotoxicity 250 fold in Vero cells and 10 fold in HeLa cells. The KDEL sequence is known to be an ER retrieval signal (17). Isolated RTA is also cytotoxic to cell lines. Because it lacks the B chain lectin, binding and subsequent uptake is very poor. Casellas et al (18) have shown for a variety of cell lines that RTA is 10,000 to 100,000 times less toxic than ricin. Presumably RTA is taken up by fluid phase mechanisms in the absence of the B chain. In addition to ricin, the castor plant Ricinus communis contains an agglutinin called RCA (19). It is a true dimer of ricin-like heterodimers, that is a heterotetramer of Ç120,000 molecular weight of form A*2B*2 . The RCA A chain (A*) is 93% identical to ricin A chain, and the RCA B chain (B*) is 84% identical to RTB (20). The A* enzyme has been shown to inactivate ribosomes with a kcat of 100/min compared with a rate of about 1500/min for RTA (21). The B* chain from RCA, unlike RTB, is univalent for galactosides (22); RCA’s agglutinating activity arise from its dimeric structure. Ebulin is a heterodimeric RIP (23). It acts as a lectin and binds galactosides. The ebulin A chain has a mass of 26 kD, roughly 6 kD smaller than RTA, yet it exhibits similar N-glycosidase rates to ricin. Ebulin is at least 10,000 times less toxic to cultured human epithelial cells than ricin, and is non-toxic to mice. In this paper we test the cytotoxicity of several native toxins, ricin point mutations, and ricin constructs against cultured human T cells. The first goal is to test if cytotoxicity is correlated with enzyme activity for ricin mutants. Once this relationship is established, it can be used to test the relative cytotoxicities of other toxins, where cell binding capacities and differences in enzyme activity are mingled. It can also be used to assess the effects of larger alterations to the toxin structure including those designed to effect membrane translocation.
which has been described previously (27). Holotoxins were preincubated for 5 minutes at 37 7C in 5 % BME to reduce the bond between A and B chains.
MATERIALS AND METHODS
Cathepsin B and D inhibition. Pepstatin A inhibits the endosomal protease cathepsin D, and leupeptin inhibits cathepsin B. To test the effects of protease inhibition on toxin activity, HSB-2 cells in log phase were resuspended in RPMI w/o methionine, containing either 0.73mM pepstatin A from Sigma (St. Louis, MO), 10mM leupeptin from Boehringer Mannheim (Indianapolis, IN), or no inhibitor. A cytotoxicity assay was performed as described above with each toxin diluted into RPMI with or without cathepsin inhibitors.
Protein expression and enzyme activity. Ricin was isolated from castor seeds as described previously (8), and fractionated into A and B chains as described by Olsnes and Pihl (24). Recombinant RTA, the Y80F, N209S, and L133N mutants, and the RTArDPro mutant were expressed and purified as for Ready et al (9). RTArKDEL was prepared as described by Wales et al (16). PAP was purified by the method of Irvin (25). Ebulin was isolated by the methods described in Girbes et al (23). RCA-A chain was cloned and expressed in pKKRCAA as described by O’Hare et al (26). Toxin enzyme activity was measured by the ability to inhibit protein synthesis in a system using ribosomes from Artemia salina
Reassociation of ricin A- and B-chain. Equimolar amounts of freshly purified recombinant RTA, or RTA mutants and plant-derived RTB (typically, each ú 500 mg) were mixed and concentrated to ú 0.5 ml using a Centricon-10 (Amicon, Inc., Beverly, MA) at 47C. Next, 2 ml 10 mM sodium phosphate, pH 7.0, containing 4 mM lactose and 1 mM EDTA (a slight modification of the reassociation buffer originally proposed by Olsnes et al. (19)) was added, and concentrated again. After 3 times buffer exchange, the solution was left at 47C. Association was monitored by gel electrophoresis. Cytotoxicity assays. HSB-2 cells obtained from American Type Culture Collection (Bethesda, MD). They are grown in RPMI 1640 (//0 Met) supplemented with 2mM L-glutamine, 1 mM Na-pyruvate, 10% FCS; all from Gibco BRL (Grand Island, NY). HSB2 cells in logphase growth were adjusted to 11106 cells/ml in RPMI 1640 without Met. 0.1 ml was dispensed to wells in 96 well plates (Costar). 100ml of each toxin dilution was added to triplicate wells and incubated 20 h at 37 7C. Wells were pulsed with 100ml of 35S-Met (NEN, Boston MA) at 1 mCi/ml for 4 h at 37 7C. The plates were harvested using a Brandel Harvester (model M-24) onto Whatman glass-fiber filter paper. Filters were counted in Bray’s Fluor using an LKB scintillation counter (LS 5000E). Protein synthesis is expressed as a percentage of control incorporation without toxin. L133N RTA. Recombinant wild-type RTA serving as doublestranded template for the site-directed mutagenesis of L133N RTA. Protein was expressed from the pUTA plasmid as described by Ready et al. (9). Mutagenesis to generate Leu133Asn (L133N) RTA was performed using the uracil-template method (28), as specified by Ready et al. (9). MBP and GST fusions. The pMAL Expression System was obtained from New England Biolabs (Beverly, MA). The GST Gene Fusion System was purchased from Pharmacia Biotech (Uppsala, Sweden). PCR was utilized to introduce RTA DNA sequences for constructing a fusion protein of RTA to the maltose binding protein (MBP) and glutathione S-transferase (GST) using primers and Taq DNA Polymerase from Applied Biosystems (Foster City CA). The pUTA plasmid (9), which contains the coding region of RTA, was used as a template for PCR. A PCR fragment of 0.85 kilo base pair was ligated into the XmnI and BamHI sites of pMAL-c2 plasmid. A PCR product using 5* primer 2 and 3* primer was digested with BamHI and ligated into the BamHI site of pGEX-2T plasmid. The ligation mixture was then transformed into Escherichia coli TB1 (araD(lac proAB) rpsL (F80_lacZDM15) hsdR) by CaCl2 method and transformants plated on YT medium containing 100 mg/ml ampicillin. TB1 Escherichia coli cells containing pMAL-RTA or pGEX-RTA were grown at YT medium containing ampicillin at 37 7C and induced with 0.3 mM isopropyl-1-thio-b-D- galactopyranoside. Bacteria were harvested as described by Ready (9). The fusion protein, MBP-RTA, was isolated by affinity chromatography as described by Guan et al. (29). The fusion protein, GST-RTA, was isolated as described by Smith and Johnson (30).
RESULTS Since inhibition of protein synthesis is the indirect measure of toxin translocation it is important to know
638
AID
BBRC 9207
/
695c$$$182
08-12-98 21:55:10
bbrcg
AP: BBRC
Vol. 249, No. 3, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS TABLE 1
Summary of Toxicity Data for RIPs and Constructs
Protein
IC50 for ribosomes (nM)
IC50 as holotoxin (pM)
IC50 for A-chain (nM)
wt RTA RCA PAP Ebulin KDEL RTA DPro-RTA MBP-RTA GST-RTA Y80F-RTA N209S-RTA L133N-RTA
0.3 2.5 0.3 0.3 ND 4.5 0.3 0.4 3 0.6 0.2
1.1 68 ND 315 ND 11 1.5 4.4 7.2 4.3 ND
260 560 320 ND 66 56 800 975 1930 900 186
Note. Column 1 lists the RIP toxin or construct, column 2 lists the IC50 against A. salina ribosomes, column 3 lists the IC50 for fluid phase cytotoxicity of the toxin A chain against T cells, and column 4 lists the IC50 of the holotoxin, A and B chains, against T cells.
the inherent rate of enzyme action for the various proteins in this study. The kinetics of several proteins used in this study have been previously characterized against ribosomes; novel constructs were tested as described in Materials and Methods. Table 1, column 2, summarizes the enzyme activity against ribosomes for the constructs used in this paper. IC50 is the concentration of toxin inhibiting 50% of ribosome activity under the assay conditions. Native ricin, isolated from Ricinus communis seeds, shows an IC50 of 0.3 nM and most of the other enzymes and mutant ricin constructs are within an order of magnitude of this value. Translocation of toxins into the cytoplasm was measured indirectly by monitoring cytotoxicity against HSB-2 T cells. Two basic cytotoxicity assays were carried out, one using holotoxins containing enzyme A chains and lectin B chains, and one using A chains alone, taken up by fluid phase mechanisms. Figure 1 shows dose response curves measuring the cytotoxicity of several native holotoxins, including ricin, RCA-I, ebulin, and ricin reconstituted from plant B chain and recombinant A chain, against HSB-2 T cells. Similar curves, not shown, were generated for ricins made by reassociating ricin B chain from plants with point mutants of RTA. Dose responses were also measured for constructs made from plant RTB and RTAs which have been modified at the termini, including fusion proteins. The characteristic value measuring cytotoxicity is the IC50 value, the concentration of toxin reducing protein synthesis in the T cells by 50% under the assay conditions; these values are listed in Table 1, column 3). Gel electrophoresis revealed that reassociation of RTB and various A chains is typically 50% complete but the conjugated proteins were readily isolated from free A and B chains by molecular sieving (data not shown).
The A chains of various RIPs, native, point mutants, and fusions, were also tested for fluid phase cytotoxicity against the T cells. In this assay, there is no B chain to efficiently bind the toxins to cell surfaces. Instead, the toxins are taken in accidentally during ‘‘cell drinking.’’ The IC50 values are listed in Table 1, column 4, and it is clear that fluid phase uptake is much less efficient than that mediated by lectin binding. For example, the IC50 for recombinant RTA is 260 nM, roughly 20,000 times higher than for ricin heterodimer. As shown in Table 1, the overall cytotoxicity of RTA fusions with GST and MBP was relatively normal. To test if these large constructs were being proteolyzed after endocytosis and being translocated as RTA, we carried out the fluid phase cytotoxicity assay in the presence of protease inhibitors pepstatin A and leupeptin. Those dose response curves (not shown) made it that the inhibitors had no major effect on cytotoxicity. DISCUSSION A key question for this study is the correlation of enzyme activity with cytotoxicity. That is, one would like to know how to break down observed differences in cytotoxicity into enzyme effects and cytoplasmic translation effects. We have examined the cytotoxicity of several active site point mutants for comparison with cytotoxicity; there is no reason to believe that these residues are involved in membrane binding, recognition, or translocation. For example, Tyr 80 participates in substrate recognition, stacking with the specific adenine target of the enzyme mechanism (10). The residue is invariant across the family of RIP enzymes because
FIG. 1. Dose response curves of the holotoxins against HSB-2 cells. Dose response for ricin (solid circles), RTA reassociated with B-chain (solid squares), RCA (diamonds), and ebulin (open circles).
639
AID
BBRC 9207
/
695c$$$182
08-12-98 21:55:10
bbrcg
AP: BBRC
Vol. 249, No. 3, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 2. Correlation of enzyme activity and fluid phase cytotoxicity assay against HSB-2.
of its role in substrate interaction. Dose response curves show that the conversion of Tyr 80 to Phe decreases overall enzyme activity 10 fold (Table 1). This loss of enzyme activity is mirrored closely in both fluid phase and B-chain mediated cytotoxicity, which are each reduced seven fold. The simplest explanation is that the mutation does not effect any aspect of translocation to the cytoplasm, and that the decrease in cytotoxicity observed is a direct reflection of the decreased enzyme activity. Figure 2 shows the relation between enzyme activity and fluid phase cytotoxicity for the native and three RTA active site mutants (Y80F, N209S, and L133N). This graph suggests a roughly linear relationship and indicates that cytotoxicity against HSB2 cells is directly correlated to enzymatic activity. As a consequence, significant deviations from this relationship for a given construct are probably indicative of translocation abnormalities. The relationship described above can now be used as a tool to analyze the behavior of various toxins. PAP is a class 1 RIP and has not evolved as part of a true heterodimeric cytotoxin. As seen in Table 1, PAP has roughly the same enzyme activity as RTA, and the same fluid phase cytotoxicity. PAP is sequestered within the plant cell wall, away from its own ribosomes and presumably enters the cytoplasm when the plant cell is compromised; there is no evidence that PAP is specialized to cross membranes (31). However, since PAP is as effective as RTA at reaching the cytoplasm from fluid phase vesicles it suggests that RTA has not evolved any special mechanisms to aid it in crossing membranes and thereby complementing RTB action.
The analysis of RCA is more complicated. The heterotetrameric lectin is reduced in activity compared to ricin (Table 1); it is 8 fold less active against ribosomes 2 fold reduced in fluid phase uptake and 60 fold lower in holotoxin activity. Saltvedt (39) had previously shown that ricin is about 20 times more cytotoxic than RCA to HeLa cells - roughly consistent with our findings against T cells. It appears that the RCA holotoxin is significantly less effective than ricin at cell uptake, even considering it has reduced enzyme activity. It is reasonable to assume that the different valency of the RCA B chain leads to reduced binding or cytoplasmic uptake. In the case of whole animal toxicity, ricin is 2000 times more cytotoxic to mice than is RCA (32). It was hypothesized that the low animal toxicity might arise from the aggregation of RCA with serum galactoproteins and its subsequent precipitation and destruction (33). Ebulin A is as active against free ribosomes as RTA, but the holotoxin is 300 fold less cytotoxic. We could not measure the fluid phase cytotoxicity in a satisfactory manner because we did not have an expression clone for ebulin A chain. All efforts to isolate the A chain from the B chain by chemical methods left about 2% holotoxin. Since the holotoxin may be 105 times as cytotoxin as the free A chain, this contaminant makes accurate measures of fluid phase uptake impossible. The reduced holotoxin cytotoxicity of ebulin probably results, like that of RCA, form deficiencies in the B chain binding. After endocytosis, the toxic proteins can move in a retrograde fashion to other membranous compartments, including the lysosomes, trans Golgi network, and ER. It has been shown that engineering the KDEL sequence onto the C terminus of RTA increases fluid phase cytotoxicity 250 fold in Vero cells and 10 fold in HeLa cells; in contrast the RTArKDEA construct acts identically to wild-type RTA (16). Presumably, the accumulation of RTArKDEL in the ER of Vero cells facilitates translocation from that vesicle. Addition of the ER retention signal to RTA increases cytotoxicity four fold in T cells (Table 1). Tagge et al (34) showed that adding the KDEL sequence to the C terminus of RTA generally increased cytotoxicity when tested against seven cell lines. The effect was more pronounced for fluid phase uptake than for endocytotic uptake, but ranged from less than a 3 fold effect for HUT102 cells to 140 fold increase against He3B cells. Immunofluorescence correlated the sensitivity to the KDEL sequenced with the presence of the KDEL receptor in the Golgi of a given cell line. The insensitivity of T cells to the KDEL sequence may arise from a poor retrieval system in this line. The RTADPro mutation removes the last six residues form the C terminus of RTA. As shown in Table 1, the deletion decreases enzyme action 15 fold. RTADPro was complexed with RTB to form a heterodimeric toxin.
640
AID
BBRC 9207
/
695c$$$182
08-12-98 21:55:10
bbrcg
AP: BBRC
Vol. 249, No. 3, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
The deletion maintains the Cys 259 residue required to bridge with RTB; the loss of the downstream residues had only a minor effect on the efficiency of reassociation with the B chain, allowing only 35% reassociation. This mutant holotoxin construct is 10 fold less cytotoxic to T cells than wild-type, consistent with its reduced enzyme activity. Interestingly, the RTADPro protein was four times more active than wild-type RTA in the fluid phase cytotoxicity assay (data not shown). This reproducible effect shows the C terminal residues may be involved in fluid phase membrane translocation. Their removal may unmask residues that facilitate transport. The presence of the B chain either blocks this effect, say by opening alternate routes for translation, or the increase in cytotoxicity facilitated by the B chain may swamp other more minor effects. Two constructs were made with intact proteins fused to the amino terminus of RTA, MBPrRTA and GSTrRTA. MBP (maltose binding protein) is a 62 kD sugar binding protein, and GST (glutathione S-transferase) is a 27 kD enzyme. As shown in Table 1, the added bulk of these fusions had very little effect on RTA’s ability to attack and inhibit ribosomes. It might have been anticipated that fusions would be retarded in cytotoxic activities, and that the MBP fusions would be more strongly effected than the GST construct. The larger size might have acted to reduce translocation to the cytoplasm, or the sugar binding activity of the MBP protein might have retarded translocation or caused alternative intracellular routing to occur. In fact, the increased size of the constructs had only a modest effect on cytotoxicity. Fluid phase uptake was inhibited only about 3 and 4 fold, respectively for the constructs. It was also possible to complex both fusion proteins with wild-type RTB from plants, although the efficiency of aggregation was poorer than for RTA. The presence of B chain greatly improved cytotoxicity, but again, the fusion constructs were at most 4 fold less toxic than ricin. It has been suggested that the intoxication of macrophages by ricin is dependent upon a proteolytic step (35). It may be that following cell uptake, the fusion proteins are cleaved by proteases and that cytotoxicity does not reflect the translocation of the bulky fusion proteins. Alternatively, if limited proteolytic processing is not essential, and does not occur, it suggests that RTA translocation is not compromised by the additional polypeptide sequence. This indicates that the translocated toxin can be used to carry large peptides that do not normally cross cellular membranes. A similar conclusion has recently been presented in the case of Pseudomonas exotoxin A (36). This observation may have utility for analytical scientific experiments and for the design of therapeutic agents. It may be possible to design modified toxins which lack antiribosomal activity but which can chaperone a variety of useful proteins into the cytoplasm.
ACKNOWLEDGMENTS This work was supported by Grant GM 30048 from the National Institutes of Health, Grant MCB-9601096 from the National Science Foundation, and by grants from the Foundation for Research and the Welch Foundation. J.S. thanks the German Academic Exchange Service for providing him a fellowship.
REFERENCES 1. Lord, J. M., Roberts, L. M., and Robertus, J. D. (1994) FASEB Journal 8, 201–208. 2. Endo, Y., and Tsurugi, K. (1987) J. Biol. Chem. 262, 8128–8130. 3. Stirpe, F., and Barbieri, L. (1985) FEBS Letters 195, 1–8. 4. Endo, Y., Tsurgi, K., and Lambert, J. M. (1988) Biochem. Biophys. Res. Comm. 150, 1032–1036. 5. van Deurs, B., Tonnenessen, T. I., Petersen, O. W., Sandvig, K., and Olsnes, S. (1986) J. Cell. Biol. 102, 37–47. 6. Neville, D. M., and Hudson, T. H. (1986) Ann. Rev. Biochem. 55, 195–224. 7. Robertus, J. D., and Monzingo, A. F. (1996) in Protein Toxin Structure (Parker, M., and Landes, R. G., Eds.), pp. 253–270, Austin, TX. 8. Rutenber, E., Katzin, B. J., Collins, E. J., Mlsna, D., Ernst, S. E., Ready, M. P., and Robertus, J. D. (1991) Proteins 10, 240–250. 9. Ready, M. P., Kim, Y. S., and Robertus, J. D. (1991) Proteins 10, 270–278. 10. Kim, Y. S., and Robertus, J. D. (1992) Protein Engineering 5, 775–779. 11. Frankel, A., Welsh, P., Richardson, J., and Robertus, J. D. (1990) Mol. Cell. Biol. 10, 6257–6263. 12. Monzingo, A. F., and Robertus, J. D. (1992) J. Mol. Biol. 227, 1136–1145. 13. Montfort, W., Villafranca, J. E., Monzingo, A. F., Ernst, S. R., Katzin, B., Rutenber, E., Nuyhen, H. X., Hamlin, R., and Robertus, J. D. (1987) J. Biol. Chem. 262, 5398–5403. 14. Newton, D. L., Wales, R., Richardson, P. T., Walbridge, S., Saxena, S. K., Ackerman, E. J., Roberts, L. M., Lord, J. M., and Youle, R. J. (1992) J. Biol. Chem. 267, 11917–11922. 15. Sandvig, K., and van Deurs, B. (1994) FEBS Lett. 346, 99–102. 16. Wales, R., Roberts, L. M., and Lord, J. M. (1993) J. Biol. Chem. 268, 23986–23990. 17. Munro, S., and Pelham, H. R. (1987) Cell 48, 899–907. 18. Casellas, P., Bourrie, J. P., Gros, P., and Jansen, F. K. (1984) J. Biol. Chem. 259, 9359–9364. 19. Olsnes, S., Saltvedt, E., and Pihl, A. (1974) J. Biol. Chem. 249, 803–810. 20. Roberts, L. M., Lamb, F. I., Pappin, D. J. C., and Lord, J. M. (1985) J. Biol. Chem. 260, 15682–15686. 21. Olsnes, S., Fernandez-Puentes, C., Carrasco, L., and Vasquez, D. (1975) Eur. J. Biochem. 60, 281–288. 22. Podder, S. K., Surolia, A., and Bachhawat, B. K. (1974) Eur. J. Biochem. 44, 151–160. 23. Girbes, T., Citores, L., Iglesias, R., Ferreras, J. M., Munoz, R., Rojo, M. A., Arias, F. J., Garcia, J. R., Mendez, E., and Calonge, M. (1993) J. Biol. Chem. 268, 18195–181199. 24. Olsnes, S., and Pihl, A. (1973) Biochemistry 12, 3121–3126. 25. Irvin, J. D. (1975) Arch. Biochem. Biophys. 169, 522–528. 26. O’Hare, M., Roberts, L. M., and Lord, J. M. (1992) FEBS Lett. 299, 209–212. 27. Ready, M. P., Rothe, G., Bird, S., and Robertus, J. D. (1984) Biochim. Biophys. Acta 740, 19–28.
641
AID
BBRC 9207
/
695c$$$182
08-12-98 21:55:10
bbrcg
AP: BBRC
Vol. 249, No. 3, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
28. Kunkel, T. A. (1985) Proc. Natl. Acad. Sci. 82, 488–492. 29. Guan, C., Li, P., Riggs, P. D., and Inouye, H. (1987) Gene 67, 21–30. 30. Smith, D. B., and Johnson, K. S. (1988) Gene 67, 31–40. 31. Ready, M. P., Brown, D. T., and Robertus, J. D. (1986) Proc. Nat. Acad. Sci. USA 83, 5053–5056. 32. Olsnes, S., Sandvig, K., Refsnes, K., and Pihl, A. (1976) J. Biol. Chem. 257, 3985–3992.
33. Saltvedt, E. (1976) Biochim. Biophys. Acta 451, 536–548. 34. Tagge, E., Chandler, J., Tang, B. L., Hong, W., Willingham, M. C., and Frankel, A. (1996) J. Histochem. Cytochem. 44, 159– 165. 35. Fiani, M. L., Blum, J. S., and Stahl, P. D. (1993) Arch. Biochem. Biophys. 307, 225–230. 36. Guidi-Rontani, C. (1996) Protein Engineering 9, 611–616.
642
AID
BBRC 9207
/
695c$$$182
08-12-98 21:55:10
bbrcg
AP: BBRC
249, 637–642 (1998)
RC989207
Differences in Cytotoxicity of Native and Engineered RIPs Can Be Used to Assess Their Ability to Reach the Cytoplasm Maria Svinth,* Jo¨rg Steighardt,* Raquel Hernandez,* Jung-Keun Suh,* Curtis Kelly,* Philip Day,† Michael Lord,† Tomas Girbes,‡ and Jon D. Robertus*,1 *Institute of Cellular and Molecular Biology, Department of Chemistry and Biochemistry, University of Texas, Austin, Texas 78712; †Department of Biological Sciences, University of Warwick, Coventry CV4 7AL, England; and ‡Departmento de Bioquimica y Biologia Molecular, Facultad de Ciencias, Universidad de Valladolid, 47005 Valladolid, Spain
Received July 20, 1998
Ricin is a heterodimeric cytotoxin composed of RTB, a galactose binding lectin, and RTA, an enzymatic Nglycosidase. The toxin is endocytosed, and after intracellular routing, RTA is translocated to the cytoplasm where it inactivates ribosomes resulting in a loss of host cell protein synthesis and cell death. We show for the first time that the cytotoxicity against cultured T cells by several RTA mutants is directly proportional to the enzyme activity of RTA, suggesting this is a reliable system to measure translocation effects. Large discrepancies between cytotoxicity and enzyme action for a given pair of toxins are therefore attributable to differences in cell binding, uptake, or membrane translocation. Fluid phase uptake and cytotoxicity of isolated RTA are essentially identical to that of the single chain toxin PAP. This important finding suggests that RTA, and the A chain of class 2 RIPs in general, has not evolved special translocation signals to complement the increased target cell binding facilitated by RTB. Experiments with the lectin RCA and with ebulin suggest those toxins have diminished cytotoxicity probably mediated by comparative deficiencies in B chain binding. Addition of a KDEL sequence to RTA increases fluid phase uptake, consistent with the notion that transport to the ER is important for cytotoxicity. Fusion of MBP or GST to the amino terminus of RTA has little effect on enzyme action or cytotoxicity. This result is not altered by protease inhibitors, suggesting the fusion proteins are probably not cleaved prior to translocation of the toxic A chain and implying that the toxins can carry large passenger proteins into the cytoplasm, an observation with interesting potential for analytical and therapeutic chemistry. q 1998 Academic Press
1 To whom correspondence should be addressed. Fax: (512) 4716135. E-mail: [email protected].
Ricin is a 65 kD heterodimeric cytotoxin composed of an A chain (RTA) and B chain (RTB) linked by a disulfide bridge (1). RTA is an N-glycosidase which recognizes a specific adenine within 28S rRNA (2), inactivating protein synthesis. RTA is a member of the ribosome inactivating proteins (RIPs) which also includes a number of single chain RIPs, such as pokeweed antiviral protein (PAP), trichosanthin, and gelonin (3). These proteins are all of similar size and all carry out the same N-glycosidation reaction (4). RTB is a lectin with an affinity for galactose; the bound toxin is endocytosed and may then traffic within a variety of intracellular compartments (5). RTA is released from RTB by reduction of the disulfide bond, and eventually reaches the cytoplasm where it enzymatically attacks ribosomes and kills the target cell (6). A good deal is known about the biochemistry and structure of RTA (7,8). Site directed mutagenesis (9,10,11) and X-ray analysis of substrate analogs have led to a mechanism of action for the RTA depurination reaction (12). RTB is a two-domain protein with two galactose binding sites (13). Site-directed mutations have confirmed the importance of sugar binding residues deduced from the structure (14). Details concerning toxin uptake, processing, intracellular routing, and membrane translocation are much less clear (15). Ricin is so toxic that only one, or a very few, molecules within the cytoplasm are required to kill a target cell. As a consequence, it is difficult to distinguish the fate of the intoxicating molecules from the bulk trafficking of what may be nonproductive pathways. On the other hand, reduction in cellular protein synthesis is an extremely sensitive indicator that RTA has reached the cytoplasm. As a consequence, protein synthesis may be used as an indirect measure of translocation of toxins into the cytoplasm.
637
0006-291X/98 $25.00 Copyright q 1998 by Academic Press All rights of reproduction in any form reserved.
AID
BBRC 9207
/
695c$$$181
08-12-98 21:55:10
bbrcg
AP: BBRC
Vol. 249, No. 3, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
Lord and coworkers have proposed that after cellular uptake, RTA may be routed to the endoplasmic reticulum (ER) and escape from there into the cytoplasm (16). This hypothesis is supported by their observation that attaching the tetrapeptide KDEL (Lys-Asp-GluLeu) to the C terminus of RTA (RTArKDEL) increases cytotoxicity 250 fold in Vero cells and 10 fold in HeLa cells. The KDEL sequence is known to be an ER retrieval signal (17). Isolated RTA is also cytotoxic to cell lines. Because it lacks the B chain lectin, binding and subsequent uptake is very poor. Casellas et al (18) have shown for a variety of cell lines that RTA is 10,000 to 100,000 times less toxic than ricin. Presumably RTA is taken up by fluid phase mechanisms in the absence of the B chain. In addition to ricin, the castor plant Ricinus communis contains an agglutinin called RCA (19). It is a true dimer of ricin-like heterodimers, that is a heterotetramer of Ç120,000 molecular weight of form A*2B*2 . The RCA A chain (A*) is 93% identical to ricin A chain, and the RCA B chain (B*) is 84% identical to RTB (20). The A* enzyme has been shown to inactivate ribosomes with a kcat of 100/min compared with a rate of about 1500/min for RTA (21). The B* chain from RCA, unlike RTB, is univalent for galactosides (22); RCA’s agglutinating activity arise from its dimeric structure. Ebulin is a heterodimeric RIP (23). It acts as a lectin and binds galactosides. The ebulin A chain has a mass of 26 kD, roughly 6 kD smaller than RTA, yet it exhibits similar N-glycosidase rates to ricin. Ebulin is at least 10,000 times less toxic to cultured human epithelial cells than ricin, and is non-toxic to mice. In this paper we test the cytotoxicity of several native toxins, ricin point mutations, and ricin constructs against cultured human T cells. The first goal is to test if cytotoxicity is correlated with enzyme activity for ricin mutants. Once this relationship is established, it can be used to test the relative cytotoxicities of other toxins, where cell binding capacities and differences in enzyme activity are mingled. It can also be used to assess the effects of larger alterations to the toxin structure including those designed to effect membrane translocation.
which has been described previously (27). Holotoxins were preincubated for 5 minutes at 37 7C in 5 % BME to reduce the bond between A and B chains.
MATERIALS AND METHODS
Cathepsin B and D inhibition. Pepstatin A inhibits the endosomal protease cathepsin D, and leupeptin inhibits cathepsin B. To test the effects of protease inhibition on toxin activity, HSB-2 cells in log phase were resuspended in RPMI w/o methionine, containing either 0.73mM pepstatin A from Sigma (St. Louis, MO), 10mM leupeptin from Boehringer Mannheim (Indianapolis, IN), or no inhibitor. A cytotoxicity assay was performed as described above with each toxin diluted into RPMI with or without cathepsin inhibitors.
Protein expression and enzyme activity. Ricin was isolated from castor seeds as described previously (8), and fractionated into A and B chains as described by Olsnes and Pihl (24). Recombinant RTA, the Y80F, N209S, and L133N mutants, and the RTArDPro mutant were expressed and purified as for Ready et al (9). RTArKDEL was prepared as described by Wales et al (16). PAP was purified by the method of Irvin (25). Ebulin was isolated by the methods described in Girbes et al (23). RCA-A chain was cloned and expressed in pKKRCAA as described by O’Hare et al (26). Toxin enzyme activity was measured by the ability to inhibit protein synthesis in a system using ribosomes from Artemia salina
Reassociation of ricin A- and B-chain. Equimolar amounts of freshly purified recombinant RTA, or RTA mutants and plant-derived RTB (typically, each ú 500 mg) were mixed and concentrated to ú 0.5 ml using a Centricon-10 (Amicon, Inc., Beverly, MA) at 47C. Next, 2 ml 10 mM sodium phosphate, pH 7.0, containing 4 mM lactose and 1 mM EDTA (a slight modification of the reassociation buffer originally proposed by Olsnes et al. (19)) was added, and concentrated again. After 3 times buffer exchange, the solution was left at 47C. Association was monitored by gel electrophoresis. Cytotoxicity assays. HSB-2 cells obtained from American Type Culture Collection (Bethesda, MD). They are grown in RPMI 1640 (//0 Met) supplemented with 2mM L-glutamine, 1 mM Na-pyruvate, 10% FCS; all from Gibco BRL (Grand Island, NY). HSB2 cells in logphase growth were adjusted to 11106 cells/ml in RPMI 1640 without Met. 0.1 ml was dispensed to wells in 96 well plates (Costar). 100ml of each toxin dilution was added to triplicate wells and incubated 20 h at 37 7C. Wells were pulsed with 100ml of 35S-Met (NEN, Boston MA) at 1 mCi/ml for 4 h at 37 7C. The plates were harvested using a Brandel Harvester (model M-24) onto Whatman glass-fiber filter paper. Filters were counted in Bray’s Fluor using an LKB scintillation counter (LS 5000E). Protein synthesis is expressed as a percentage of control incorporation without toxin. L133N RTA. Recombinant wild-type RTA serving as doublestranded template for the site-directed mutagenesis of L133N RTA. Protein was expressed from the pUTA plasmid as described by Ready et al. (9). Mutagenesis to generate Leu133Asn (L133N) RTA was performed using the uracil-template method (28), as specified by Ready et al. (9). MBP and GST fusions. The pMAL Expression System was obtained from New England Biolabs (Beverly, MA). The GST Gene Fusion System was purchased from Pharmacia Biotech (Uppsala, Sweden). PCR was utilized to introduce RTA DNA sequences for constructing a fusion protein of RTA to the maltose binding protein (MBP) and glutathione S-transferase (GST) using primers and Taq DNA Polymerase from Applied Biosystems (Foster City CA). The pUTA plasmid (9), which contains the coding region of RTA, was used as a template for PCR. A PCR fragment of 0.85 kilo base pair was ligated into the XmnI and BamHI sites of pMAL-c2 plasmid. A PCR product using 5* primer 2 and 3* primer was digested with BamHI and ligated into the BamHI site of pGEX-2T plasmid. The ligation mixture was then transformed into Escherichia coli TB1 (araD(lac proAB) rpsL (F80_lacZDM15) hsdR) by CaCl2 method and transformants plated on YT medium containing 100 mg/ml ampicillin. TB1 Escherichia coli cells containing pMAL-RTA or pGEX-RTA were grown at YT medium containing ampicillin at 37 7C and induced with 0.3 mM isopropyl-1-thio-b-D- galactopyranoside. Bacteria were harvested as described by Ready (9). The fusion protein, MBP-RTA, was isolated by affinity chromatography as described by Guan et al. (29). The fusion protein, GST-RTA, was isolated as described by Smith and Johnson (30).
RESULTS Since inhibition of protein synthesis is the indirect measure of toxin translocation it is important to know
638
AID
BBRC 9207
/
695c$$$182
08-12-98 21:55:10
bbrcg
AP: BBRC
Vol. 249, No. 3, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS TABLE 1
Summary of Toxicity Data for RIPs and Constructs
Protein
IC50 for ribosomes (nM)
IC50 as holotoxin (pM)
IC50 for A-chain (nM)
wt RTA RCA PAP Ebulin KDEL RTA DPro-RTA MBP-RTA GST-RTA Y80F-RTA N209S-RTA L133N-RTA
0.3 2.5 0.3 0.3 ND 4.5 0.3 0.4 3 0.6 0.2
1.1 68 ND 315 ND 11 1.5 4.4 7.2 4.3 ND
260 560 320 ND 66 56 800 975 1930 900 186
Note. Column 1 lists the RIP toxin or construct, column 2 lists the IC50 against A. salina ribosomes, column 3 lists the IC50 for fluid phase cytotoxicity of the toxin A chain against T cells, and column 4 lists the IC50 of the holotoxin, A and B chains, against T cells.
the inherent rate of enzyme action for the various proteins in this study. The kinetics of several proteins used in this study have been previously characterized against ribosomes; novel constructs were tested as described in Materials and Methods. Table 1, column 2, summarizes the enzyme activity against ribosomes for the constructs used in this paper. IC50 is the concentration of toxin inhibiting 50% of ribosome activity under the assay conditions. Native ricin, isolated from Ricinus communis seeds, shows an IC50 of 0.3 nM and most of the other enzymes and mutant ricin constructs are within an order of magnitude of this value. Translocation of toxins into the cytoplasm was measured indirectly by monitoring cytotoxicity against HSB-2 T cells. Two basic cytotoxicity assays were carried out, one using holotoxins containing enzyme A chains and lectin B chains, and one using A chains alone, taken up by fluid phase mechanisms. Figure 1 shows dose response curves measuring the cytotoxicity of several native holotoxins, including ricin, RCA-I, ebulin, and ricin reconstituted from plant B chain and recombinant A chain, against HSB-2 T cells. Similar curves, not shown, were generated for ricins made by reassociating ricin B chain from plants with point mutants of RTA. Dose responses were also measured for constructs made from plant RTB and RTAs which have been modified at the termini, including fusion proteins. The characteristic value measuring cytotoxicity is the IC50 value, the concentration of toxin reducing protein synthesis in the T cells by 50% under the assay conditions; these values are listed in Table 1, column 3). Gel electrophoresis revealed that reassociation of RTB and various A chains is typically 50% complete but the conjugated proteins were readily isolated from free A and B chains by molecular sieving (data not shown).
The A chains of various RIPs, native, point mutants, and fusions, were also tested for fluid phase cytotoxicity against the T cells. In this assay, there is no B chain to efficiently bind the toxins to cell surfaces. Instead, the toxins are taken in accidentally during ‘‘cell drinking.’’ The IC50 values are listed in Table 1, column 4, and it is clear that fluid phase uptake is much less efficient than that mediated by lectin binding. For example, the IC50 for recombinant RTA is 260 nM, roughly 20,000 times higher than for ricin heterodimer. As shown in Table 1, the overall cytotoxicity of RTA fusions with GST and MBP was relatively normal. To test if these large constructs were being proteolyzed after endocytosis and being translocated as RTA, we carried out the fluid phase cytotoxicity assay in the presence of protease inhibitors pepstatin A and leupeptin. Those dose response curves (not shown) made it that the inhibitors had no major effect on cytotoxicity. DISCUSSION A key question for this study is the correlation of enzyme activity with cytotoxicity. That is, one would like to know how to break down observed differences in cytotoxicity into enzyme effects and cytoplasmic translation effects. We have examined the cytotoxicity of several active site point mutants for comparison with cytotoxicity; there is no reason to believe that these residues are involved in membrane binding, recognition, or translocation. For example, Tyr 80 participates in substrate recognition, stacking with the specific adenine target of the enzyme mechanism (10). The residue is invariant across the family of RIP enzymes because
FIG. 1. Dose response curves of the holotoxins against HSB-2 cells. Dose response for ricin (solid circles), RTA reassociated with B-chain (solid squares), RCA (diamonds), and ebulin (open circles).
639
AID
BBRC 9207
/
695c$$$182
08-12-98 21:55:10
bbrcg
AP: BBRC
Vol. 249, No. 3, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 2. Correlation of enzyme activity and fluid phase cytotoxicity assay against HSB-2.
of its role in substrate interaction. Dose response curves show that the conversion of Tyr 80 to Phe decreases overall enzyme activity 10 fold (Table 1). This loss of enzyme activity is mirrored closely in both fluid phase and B-chain mediated cytotoxicity, which are each reduced seven fold. The simplest explanation is that the mutation does not effect any aspect of translocation to the cytoplasm, and that the decrease in cytotoxicity observed is a direct reflection of the decreased enzyme activity. Figure 2 shows the relation between enzyme activity and fluid phase cytotoxicity for the native and three RTA active site mutants (Y80F, N209S, and L133N). This graph suggests a roughly linear relationship and indicates that cytotoxicity against HSB2 cells is directly correlated to enzymatic activity. As a consequence, significant deviations from this relationship for a given construct are probably indicative of translocation abnormalities. The relationship described above can now be used as a tool to analyze the behavior of various toxins. PAP is a class 1 RIP and has not evolved as part of a true heterodimeric cytotoxin. As seen in Table 1, PAP has roughly the same enzyme activity as RTA, and the same fluid phase cytotoxicity. PAP is sequestered within the plant cell wall, away from its own ribosomes and presumably enters the cytoplasm when the plant cell is compromised; there is no evidence that PAP is specialized to cross membranes (31). However, since PAP is as effective as RTA at reaching the cytoplasm from fluid phase vesicles it suggests that RTA has not evolved any special mechanisms to aid it in crossing membranes and thereby complementing RTB action.
The analysis of RCA is more complicated. The heterotetrameric lectin is reduced in activity compared to ricin (Table 1); it is 8 fold less active against ribosomes 2 fold reduced in fluid phase uptake and 60 fold lower in holotoxin activity. Saltvedt (39) had previously shown that ricin is about 20 times more cytotoxic than RCA to HeLa cells - roughly consistent with our findings against T cells. It appears that the RCA holotoxin is significantly less effective than ricin at cell uptake, even considering it has reduced enzyme activity. It is reasonable to assume that the different valency of the RCA B chain leads to reduced binding or cytoplasmic uptake. In the case of whole animal toxicity, ricin is 2000 times more cytotoxic to mice than is RCA (32). It was hypothesized that the low animal toxicity might arise from the aggregation of RCA with serum galactoproteins and its subsequent precipitation and destruction (33). Ebulin A is as active against free ribosomes as RTA, but the holotoxin is 300 fold less cytotoxic. We could not measure the fluid phase cytotoxicity in a satisfactory manner because we did not have an expression clone for ebulin A chain. All efforts to isolate the A chain from the B chain by chemical methods left about 2% holotoxin. Since the holotoxin may be 105 times as cytotoxin as the free A chain, this contaminant makes accurate measures of fluid phase uptake impossible. The reduced holotoxin cytotoxicity of ebulin probably results, like that of RCA, form deficiencies in the B chain binding. After endocytosis, the toxic proteins can move in a retrograde fashion to other membranous compartments, including the lysosomes, trans Golgi network, and ER. It has been shown that engineering the KDEL sequence onto the C terminus of RTA increases fluid phase cytotoxicity 250 fold in Vero cells and 10 fold in HeLa cells; in contrast the RTArKDEA construct acts identically to wild-type RTA (16). Presumably, the accumulation of RTArKDEL in the ER of Vero cells facilitates translocation from that vesicle. Addition of the ER retention signal to RTA increases cytotoxicity four fold in T cells (Table 1). Tagge et al (34) showed that adding the KDEL sequence to the C terminus of RTA generally increased cytotoxicity when tested against seven cell lines. The effect was more pronounced for fluid phase uptake than for endocytotic uptake, but ranged from less than a 3 fold effect for HUT102 cells to 140 fold increase against He3B cells. Immunofluorescence correlated the sensitivity to the KDEL sequenced with the presence of the KDEL receptor in the Golgi of a given cell line. The insensitivity of T cells to the KDEL sequence may arise from a poor retrieval system in this line. The RTADPro mutation removes the last six residues form the C terminus of RTA. As shown in Table 1, the deletion decreases enzyme action 15 fold. RTADPro was complexed with RTB to form a heterodimeric toxin.
640
AID
BBRC 9207
/
695c$$$182
08-12-98 21:55:10
bbrcg
AP: BBRC
Vol. 249, No. 3, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
The deletion maintains the Cys 259 residue required to bridge with RTB; the loss of the downstream residues had only a minor effect on the efficiency of reassociation with the B chain, allowing only 35% reassociation. This mutant holotoxin construct is 10 fold less cytotoxic to T cells than wild-type, consistent with its reduced enzyme activity. Interestingly, the RTADPro protein was four times more active than wild-type RTA in the fluid phase cytotoxicity assay (data not shown). This reproducible effect shows the C terminal residues may be involved in fluid phase membrane translocation. Their removal may unmask residues that facilitate transport. The presence of the B chain either blocks this effect, say by opening alternate routes for translation, or the increase in cytotoxicity facilitated by the B chain may swamp other more minor effects. Two constructs were made with intact proteins fused to the amino terminus of RTA, MBPrRTA and GSTrRTA. MBP (maltose binding protein) is a 62 kD sugar binding protein, and GST (glutathione S-transferase) is a 27 kD enzyme. As shown in Table 1, the added bulk of these fusions had very little effect on RTA’s ability to attack and inhibit ribosomes. It might have been anticipated that fusions would be retarded in cytotoxic activities, and that the MBP fusions would be more strongly effected than the GST construct. The larger size might have acted to reduce translocation to the cytoplasm, or the sugar binding activity of the MBP protein might have retarded translocation or caused alternative intracellular routing to occur. In fact, the increased size of the constructs had only a modest effect on cytotoxicity. Fluid phase uptake was inhibited only about 3 and 4 fold, respectively for the constructs. It was also possible to complex both fusion proteins with wild-type RTB from plants, although the efficiency of aggregation was poorer than for RTA. The presence of B chain greatly improved cytotoxicity, but again, the fusion constructs were at most 4 fold less toxic than ricin. It has been suggested that the intoxication of macrophages by ricin is dependent upon a proteolytic step (35). It may be that following cell uptake, the fusion proteins are cleaved by proteases and that cytotoxicity does not reflect the translocation of the bulky fusion proteins. Alternatively, if limited proteolytic processing is not essential, and does not occur, it suggests that RTA translocation is not compromised by the additional polypeptide sequence. This indicates that the translocated toxin can be used to carry large peptides that do not normally cross cellular membranes. A similar conclusion has recently been presented in the case of Pseudomonas exotoxin A (36). This observation may have utility for analytical scientific experiments and for the design of therapeutic agents. It may be possible to design modified toxins which lack antiribosomal activity but which can chaperone a variety of useful proteins into the cytoplasm.
ACKNOWLEDGMENTS This work was supported by Grant GM 30048 from the National Institutes of Health, Grant MCB-9601096 from the National Science Foundation, and by grants from the Foundation for Research and the Welch Foundation. J.S. thanks the German Academic Exchange Service for providing him a fellowship.
REFERENCES 1. Lord, J. M., Roberts, L. M., and Robertus, J. D. (1994) FASEB Journal 8, 201–208. 2. Endo, Y., and Tsurugi, K. (1987) J. Biol. Chem. 262, 8128–8130. 3. Stirpe, F., and Barbieri, L. (1985) FEBS Letters 195, 1–8. 4. Endo, Y., Tsurgi, K., and Lambert, J. M. (1988) Biochem. Biophys. Res. Comm. 150, 1032–1036. 5. van Deurs, B., Tonnenessen, T. I., Petersen, O. W., Sandvig, K., and Olsnes, S. (1986) J. Cell. Biol. 102, 37–47. 6. Neville, D. M., and Hudson, T. H. (1986) Ann. Rev. Biochem. 55, 195–224. 7. Robertus, J. D., and Monzingo, A. F. (1996) in Protein Toxin Structure (Parker, M., and Landes, R. G., Eds.), pp. 253–270, Austin, TX. 8. Rutenber, E., Katzin, B. J., Collins, E. J., Mlsna, D., Ernst, S. E., Ready, M. P., and Robertus, J. D. (1991) Proteins 10, 240–250. 9. Ready, M. P., Kim, Y. S., and Robertus, J. D. (1991) Proteins 10, 270–278. 10. Kim, Y. S., and Robertus, J. D. (1992) Protein Engineering 5, 775–779. 11. Frankel, A., Welsh, P., Richardson, J., and Robertus, J. D. (1990) Mol. Cell. Biol. 10, 6257–6263. 12. Monzingo, A. F., and Robertus, J. D. (1992) J. Mol. Biol. 227, 1136–1145. 13. Montfort, W., Villafranca, J. E., Monzingo, A. F., Ernst, S. R., Katzin, B., Rutenber, E., Nuyhen, H. X., Hamlin, R., and Robertus, J. D. (1987) J. Biol. Chem. 262, 5398–5403. 14. Newton, D. L., Wales, R., Richardson, P. T., Walbridge, S., Saxena, S. K., Ackerman, E. J., Roberts, L. M., Lord, J. M., and Youle, R. J. (1992) J. Biol. Chem. 267, 11917–11922. 15. Sandvig, K., and van Deurs, B. (1994) FEBS Lett. 346, 99–102. 16. Wales, R., Roberts, L. M., and Lord, J. M. (1993) J. Biol. Chem. 268, 23986–23990. 17. Munro, S., and Pelham, H. R. (1987) Cell 48, 899–907. 18. Casellas, P., Bourrie, J. P., Gros, P., and Jansen, F. K. (1984) J. Biol. Chem. 259, 9359–9364. 19. Olsnes, S., Saltvedt, E., and Pihl, A. (1974) J. Biol. Chem. 249, 803–810. 20. Roberts, L. M., Lamb, F. I., Pappin, D. J. C., and Lord, J. M. (1985) J. Biol. Chem. 260, 15682–15686. 21. Olsnes, S., Fernandez-Puentes, C., Carrasco, L., and Vasquez, D. (1975) Eur. J. Biochem. 60, 281–288. 22. Podder, S. K., Surolia, A., and Bachhawat, B. K. (1974) Eur. J. Biochem. 44, 151–160. 23. Girbes, T., Citores, L., Iglesias, R., Ferreras, J. M., Munoz, R., Rojo, M. A., Arias, F. J., Garcia, J. R., Mendez, E., and Calonge, M. (1993) J. Biol. Chem. 268, 18195–181199. 24. Olsnes, S., and Pihl, A. (1973) Biochemistry 12, 3121–3126. 25. Irvin, J. D. (1975) Arch. Biochem. Biophys. 169, 522–528. 26. O’Hare, M., Roberts, L. M., and Lord, J. M. (1992) FEBS Lett. 299, 209–212. 27. Ready, M. P., Rothe, G., Bird, S., and Robertus, J. D. (1984) Biochim. Biophys. Acta 740, 19–28.
641
AID
BBRC 9207
/
695c$$$182
08-12-98 21:55:10
bbrcg
AP: BBRC
Vol. 249, No. 3, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
28. Kunkel, T. A. (1985) Proc. Natl. Acad. Sci. 82, 488–492. 29. Guan, C., Li, P., Riggs, P. D., and Inouye, H. (1987) Gene 67, 21–30. 30. Smith, D. B., and Johnson, K. S. (1988) Gene 67, 31–40. 31. Ready, M. P., Brown, D. T., and Robertus, J. D. (1986) Proc. Nat. Acad. Sci. USA 83, 5053–5056. 32. Olsnes, S., Sandvig, K., Refsnes, K., and Pihl, A. (1976) J. Biol. Chem. 257, 3985–3992.
33. Saltvedt, E. (1976) Biochim. Biophys. Acta 451, 536–548. 34. Tagge, E., Chandler, J., Tang, B. L., Hong, W., Willingham, M. C., and Frankel, A. (1996) J. Histochem. Cytochem. 44, 159– 165. 35. Fiani, M. L., Blum, J. S., and Stahl, P. D. (1993) Arch. Biochem. Biophys. 307, 225–230. 36. Guidi-Rontani, C. (1996) Protein Engineering 9, 611–616.
642
AID
BBRC 9207
/
695c$$$182
08-12-98 21:55:10
bbrcg
AP: BBRC